EON is a modern, component-based architecture that developers can use to build robust decision-support systems that reason about guideline- directed care. The system comprises a number of modules, including (1) problem solvers that plan medical therapy and determine whether patients are eligible for particular protocols, (2) a database mediator that handles all queries posed by the problem solvers to return point patient data or temporal abstractions of those data in a transparent fashion, and (3) a shared knowledge based of clinical guidelines and basic medical knowledge. EON is not a standalone program, but rather constitutes middleware that is intended to be embedded within a given clinical information system. In this competing renewal application, we propose to develop and evaluate new computational methods to improve the EON architecture in a variety of ways to make the components more useful, explainable, and maintainable. Our specific aims include (1) creation of a set of interactive interfaces that will allow clinicians to visualize the behavior of the EON components and the data that they process, (2) extension of the capabilities of the EON database mediator to answer more complex time-dependent queries, (3) reformulation of the knowledge bases on which the EON components operate as CORBA-compliant servers, and (4) extensions to the problem-solving components that automate specific tasks related to guideline-directed therapy. In particular, we will develop CORBA-based mapping objects that allow problem solvers and knowledge bases of clinical protocols to interoperate over the Internet. We will evaluate the usability of the new EON components both within our laboratory and at another clinical site outside of Stanford. We will use EON to encode and execute protocols for breast cancer, AIDS, hypertension, and diabetes. Our goals are to develop EON into a set of middleware components that are sufficiently robust to handle the complexities of an extremely wide range of clinical protocols and guidelines, to demonstrate the effectiveness of a comprehensive decision-support system that is compatible with modern component-based software architectures, to provide better insight into the basis for EON's therapy recommendations, and to refine EON's knowledge-acquisition methodology so that clinicians can author and maintain electronic knowledge bases of protocols and guidelines more confidently and independently.